Computer modeling helps to improve the quality and microbiological safety of food
January 22, 2014
By Elton Alisson
Agência FAPESP – The foodstuff industries of several countries have utilized a new tool to improve the microbiological safety and quality of their products. Predictive microbiology is a software-based system that uses mathematical models and statistics to predict the behavior of microorganisms in fresh and processed food.
The new method is based on the principle that the ability of bacteria and fungi to multiply in food depends on the properties of the product, such as its composition, acidity, humidity, salt levels and antimicrobials present, in addition to the temperature conditions, relative humidity and atmosphere in which it is maintained. In this manner, the effect of each of these factors can be calculated mathematically using different predictive models.
Because of the set of benefits it offers, this method has been replacing traditional forms of assessing food contamination risks, which are primarily conducted through microbiological analyses in laboratories. In addition to being expensive, these traditional methods have limited value because the results are only valid for samples that are directly evaluated, according to researchers in the field.
“Because microorganisms are not distributed uniformly in food, it is necessary to conduct microbiological analyses on many samples of the product to conclude whether it is safe for consumption,” said Bernadette Dora Gombossy de Melo Franco, a professor at the School of Pharmaceutical Sciences at Universidade de São Paulo (USP).
“The predictive model takes into consideration the statistical aspects of variability and the uncertainty of measurements, coupled with the imprecision of laboratory methods of analysis, to reach these conclusions,” explained Franco, who is the coordinator of the Food Research Center (FoRC), one of the Research, Innovation and Dissemination Centers (CEPID).
According to Franco, with modeling, it is possible to forecast how long a product will remain fresh (shelf life) and what can be done technologically to improve this.
Using mathematical calculations, one can evaluate the effects on shelf life if, for example, hot dogs that should be refrigerated at 12°C, as the factories recommend, are stored at 15°C or 18°C instead. These calculations can also be used to determine whether another preservative or a new processing technology will affect the shelf life.
The method can also be used to make quantitative evaluations of the health risk to the consumer from the farm to the fork, depending on the pathogenic microorganism considered, the production and commercial conditions, and the type of consumption, according to Franco.
“In Brazil, there is a general underreporting of food-borne infirmities caused by contaminating microorganisms, and we know that this problem is much greater than we imagined,” said the researcher.
“Climatic factors and the long distances between production and consumption sites favor these occurrences, particularly if the transport conditions are inadequate,” she stressed.
ESPCA on the theme
To promote broader usage and application of the new method in Brazil, the Department of Food and Nutrition held a São Paulo School of Advanced Science – Advances in predictive modeling and quantitative microbiological risk assessment of foods, from October 28 to November 5, 2013.
Coordinated by Franco and conducted under the auspices of the São Paulo School of Advanced Sciences (ESPCA) – a FAPESP modality that funds short courses in different areas of knowledge – the event brought together dozens of students and 17 renowned researchers from the United States, Denmark, Malta, Spain, the Netherlands, Belgium, Australia, France, Greece and Brazil, all specialists in predictive microbiology. Among the specialists was Paw Dalgaard, a professor at Denmark Technical University.
Dalgaard and his colleagues from his research group in predictive microbiology developed software to forecast the effect of constant and variable temperature conditions on the expiration date of seasoned tropical fish and seafood and on the growth of bacteria, such as Photobacterium phosphoreum and Shewanella puftefaciens, in these products.
Available on the Internet at no cost, the software can help the fish industry to forecast pathogenic bacterial growth in fresh fish, for example, affirmed the researcher.
“In practice, fresh and lightly preserved foods, such as fish and seafood, can be conserved at temperatures between 0°C and 15°C, but in tropical regions they are often exposed to temperatures above this range, facilitating the growth of bacteria in the product,” said Dalgaard.
In Denmark, through the reduction and control of the commercial storage temperatures used by the country’s fisheries industry based on microbiological models, it was possible to eliminate the use of salt as a preservative in these products, explained the researcher.
It took many years, however, for the Danish fisheries industry to understand how predictive modeling could be used to reduce the risk of contamination of fish, noted Dalgaard. “We had to change the entire preservation culture that was followed for decades,” he affirmed.
Another renowned researcher participating in the event was József Baranyi, of the Food Research Institute in Norwich in the United Kingdom.
Considered the father of predictive modeling, Baranyi developed and manages the Combined dataBase for predictive microbiology (ComBase), an immense database on the behavior of microorganisms under different food conditions> He also manages a collection of predictive models available on the Internet for varied applications in the food industry, along with researchers from Australia and the United States.
Delays in Brazil
In addition to European countries, including Denmark and the United Kingdom, the United States has already adopted predictive modeling.
In the 1980s, for example, the U.S. Department of Agriculture (USDA) developed a pioneering program for modeling pathogens in foods, termed the Pathogen Modeling Program (PMP), which includes growth, survival and deactivation models and is widely used by several countries.
Researcher Robert Buchanan, of the University of Maryland’s Food Safety Center, who participated in the development of PMP, was one of the speakers at the ESPCA on Predictive Modeling.
In Brazil, one of the world’s largest exporters of food, there are few researchers working in this area, noted the researchers participating in the event.
“The research community in this area must be expanded so that we do not fall behind what the world is doing today in terms of new techniques to reduce the consequences of microbial contamination in food,” stated Franco. “For this reason, holding this event in Brazil was important.”
Franco recently initiated a research project in partnership with colleagues from Denmark Technical University.
Held with FAPESP funding under the auspices of the agreement signed by the foundation and the Danish Council for Strategic Research (DCSR), the project aims to develop a predictive model of the cross contamination of meat products with Salmonella and Listeria monocytogenes during slaughtering and product processing.
The results of the project will allow the industry and commercial establishments to develop control systems to minimize the risks of product contamination, predict the researchers.
Because of the tropical climate in Brazil, food in the country has a very high load of bacterial spores with high thermal resistance, said researchers participating in the event.
For this reason, according to the researchers, Brazil must develop models of chemical disinfection for Brazilian food industries to determine the types and quantities of sanitizer necessary to disinfect production lines.
“Many of these bacterial spores form biofilms or coat the walls of the processing equipment and are hard to eliminate,” explained Pilar Rodriguez de Massaguer, a retired professor, formerly at the School of Food Engineering (FEA), and a collaborating researcher with the School of Chemical Engineering (FEQ) at Universidade Estadual de Campinas (Unicamp).
“If there were predictive models of disinfection, the industries could better understand these microorganisms and know which products would help to eliminate them,” she noted.
Because these bacterial spores survive and can eventually leave the foodstuff industries, there is also a need for predictive models that indicate how long the microorganisms that come into contact with food in the factory will survive on supermarket shelves before germinating, growing and contaminating the product, indicated Massaguer.
Within supermarket environments, according to Massaguer, there is also a need to model, for example, the effect of temperature variations on the preservation of refrigerated food.
“The temperature in the refrigeration chambers of supermarkets in Brazil varies between 10°C and 15°C and never stays at 4°C [the preservation temperature for refrigerated food],” said Massaguer. “Evaluating the effect of this variable temperature in limiting the shelf life of several meat based products and fruit and milk produced in Brazil is a must.”
According to Massaguer, the Brazilian foodstuff industry uses the predictive model only to solve specific problems by a contract with researchers and specialists in the area. In addition, at present, there is no industrial investment in the development of software to conduct predictive microbiology, she noted.
“Predictive modeling is a multidisciplinary area that requires the involvement of not only food engineers but also mathematical and statistical engineers. Such collaboration is still in its early stages in Brazil, which is one of the reasons why it [predictive modeling] is seldom used by the Brazilian food industry,” affirmed Massaguer.